Article Text
PDF
Abstract
When COVID-19 hit India, a qualitative research study had been underway the southern state of Kerala, to understand the perspectives of the front-line health workers and the Kattunayakan tribal community towards health service utilisation. This community is relatively underserved, and a great deal of our emphasis was on understanding health system barriers experienced on both demand and supply side. COVID-19 showed us that these barriers pertain not just to heath systems, but also to the conduct of health research. We completed fieldwork in one hamlet before lockdowns were announced and changed our fieldwork approach for the remaining two different hamlets. The main change was a shift to the use of mobile telephony for fieldwork. This technological shift necessitated substantial changes in the design of fieldwork, the scope of our inquiry, as well as the composition and power dynamics within our team. First, adjusting to technology-driven fieldwork posed restrictions but also enhanced the agency and comfort of participants in some ways. Study design changes attributable to COVID-19 restrictions were brought about, but also gave us critical insight into the impact of COVID-19 and related outbreaks. There was de fact greater reliance on community researchers, which meant we ceded control to the community itself, upsetting typical research power dynamics, which can be quite top-down. We present these methodological reflections for wider consideration.
- COVID-19
- health services research
- qualitative study
Data availability statement
The authors confirm that the data supporting the findings of this study are available within the article.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
Summary box
COVID-19 has required transformation of field research approaches, but a switch to digital formats is not possible for all populations.
Our qualitative research examined experiences of health services utilisation by indigenous Kattunayakan communities in southern India employed telephony given lockdown restrictions during COVID-19 .
This technological shift necessitated substantial changes in the design of fieldwork, the scope of our inquiry, as well as the composition and power dynamics within our team and in relation to the community.
Population groups that are underserved by health systems should not be under-represented in health research, even if there are unique trade-offs in reaching them.
Introduction
The COVID-19 pandemic has seized almost every country in the globe; in India, the number of cases has been steadily rising across waves. The south Indian state of Kerala was the first to announce a case in early 2020; the state’s response was hailed but has also led to some circumspection about what could have been done better.1 2 In a recent global seminar, coordinated efforts in the northern district of Wayanad were credited with saving lives and ensuring economic stability.3
At the time the pandemic broke out, that is, early January, our team had qualitative research underway in this district. Our study sought to understand the perspectives of front-line health workers as well as Kattu-nayakan (tribal) communities in this district of Kerala with respect to health service utilisation. It involved focus group discussions (FGDs) and group interviews with Kattunayakan tribal communities living in hamlets of this area, as well as front-line and facility level health workers tasked with serving them. We had been constructing case studies of these stakeholders from criterion-sampled hamlets, seeking to understand enablers and barriers to care from both the demand and supply side. As we completed one case study and were preparing to begin another, COVID-19 lockdowns were announced.
As research teams around us struggled to keep their research ongoing, we felt relief that our study, employed qualitative methods specifically designed to be sensitive to context.4 However, we were also mindful that our in-person fieldwork may eventually have to be replaced with other, digital or remotely administered methods. As we were working with participants from tribal populations for whom internet access was expensive and irregular, and phone ownership uneven, we anticipated that the most appropriate method would likely have to be telephonic interviews. We were, of course aware of the gamut of concerns when shifting from face-to-face interactions to digital media for data collection,5 which include problems with exclusion in recruitment, difficulties in maintaining flow and ensuring clarity during FGDs, as well as technical/accessibility-related constraints. There were also broader questions of ‘automating inequality’, that is stereotyping, penalising or excluding certain populations in the application of technology5 (including health research fieldwork), which we did not want to contribute towards.
Drawing on this and other literature, as well as inputs from other teams similarly seeking to continue fieldwork in virtuo, we amended our study protocol so as to enable telephonic group and individual interveiws adapting existing aspects of our fieldwork. This paper is a reflection on what we proposed to do, what we ended up doing, and the trade-offs that resulted.
Initial study design
As per our original protocol, approved by the ethics committee of the George Institute for Global Health (007/2018), we used district programme data to enlist areas where Kerala’s newly launched Oorumithram or hamlet ASHA programme had been put in place to identify hamlets where (1) there was no hamlet ASHA; (2) there was a hamlet ASHA working with multiple tribal groups and (3) there was a hamlet ASHA working only with Kattunayakans. About half of the total list of hamlets met our criteria and we had intended to create age-sex disaggregated groups at each site. Alongside this, a Kattunayakan community researcher joined our team; she had been introduced to us in earlier phases of research, including its design and had agreed to facilitate connections, provide us with critical contextual information and participate in the conduct of fieldwork and analysis as a member of our team. We had earlier secured permissions from the district medical officer, district administration, as well as the state tribal welfare department, the state forestry department, as well as the state health department.
Just as we had completed the fieldwork at our first site (which adjoined our community researcher’s place of residence), a nationwide lockdown was declared (in late March 2020). Since some of our permissions were time-bound, and our project was due to end in the middle of the year, we checked in on permissions and revisited our timeframe. We were informed that field permissions were not being granted (nor were extensions, which we would likely need). It became increasingly clear that we would have to rely on virtual or telephonic means of completing our fieldwork. At this time, there emerged a profusion of documentation, in blogs and peer reviewed articles, related to the conduct of fieldwork during COVID-19.4 6–11 We reviewed these, while also speaking to peers conducting research, as well as collaborators and key informants from our previous fieldwork to come up with a revised ‘non-field’ plan of action.
Revising our approach
Our first step was to update our list of protocols for the community gathering and the precautionary measures in the state to ensure that we were in adherence during the proposed digital field work. In one case, we were informed by community members that abidance of these protocols was not strict and that we could enter field sites and carry out fieldwork; however, erring on the side of compliance and abundant precaution, we made the decision to continue with the revised process and sought an amendment to our ethics protocol with a number of changes.
While a number of resources on COVID-19 fieldwork were emerging at the time regarding virtual methods (zoom interviews, consent using Whatsapp, etc), these resources were not relevant to our situation. Phone ownership in our community was generally low, phone use and access highly variable, and network connectivity extremely poor. It became quite clear that telephony was the most viable medium—even with all its constraints—to conduct fieldwork during COVID-19 lockdowns. Since most of the resources did not pertain to this, we relied on the few relevant resources,7 and we added our own experiences, as well as insights and advice from our networks within the location and from the community to move forward.
A major proposed change was to the remit of the fieldwork—while we had already shifted from FGDs to group interviews to accommodate smaller groups of participants meeting our criteria, we opened up the possibility of having to conduct distanced group interviews and possibly individual interviews. This was due to COVID-19 social distancing restrictions and further anticipated difficulties in identifying eligible participants. We made the decision that the number of interviews would be decided based on looking at availability of participants in this changed context, with data saturation being a secondary consideration.
Second, since we knew now that our interviews would likely take place on phones and that phone ownership in our areas was unlikely to be ubiquitous, we sought permission from our ethics committee for recorded verbal consent as an option in addition to written informed consent, again to minimise interaction and risk associated with COVID-19 (in terms of handling paper, pens and consent forms, among others). Participant Information Sheets would continue to be shared, with care taken to ensure that the process of printing and sharing these would not introduce exposure risk.
Perhaps the most important change was with respect to our team itself, which changed the very nature of our data collection thereafter. The community researcher in our team had personal health reasons, in addition to lockdown travel constraints that precluded her from continuing to work with us.
Our revised data collection plan relied on one community researcher per location of fieldwork, that is, given COVID-19 mobility restrictions, we sought to identify a separate community researcher in each hamlet where we wished to gather data. A key criterion for the post of community researcher was the possession of a phone and charger (we compensated community researchers for their phone use as part of their stipend for assisting us on this project). We sought nominations from front-line health workers of the respective hamlets for selecting the community researchers, which were corroborated by one other worker (the Junior Public Health Nurse or tribal promoter) in each area. We then spoke to nominees one at a time to inform them of the requirements, gauge their interest and suitability, and finalise their selection. In one case, we had identified a community researcher but were not able to sustain communication with her and therefore decided to select another hamlet. We got confirmation from two selected individuals, a male pursuing higher education and a married female who had completed high school education, who we then individually oriented and trained via phone regarding our study scope and methodology.
Given this change in our process, we realised we would have less time to work with each community researcher, on the other we would rely on them all the more. Task-shifting happened from the investigators to the community researchers when the non-field plan was rolled out. This included key aspects of fieldwork. For instance, community researchers were entrusted with creating groups for interviews and they were scheduling the discussions and interviews of participants, by keeping the investigator informed over the phone. They were responsible for selecting a private location that would also allow for social distancing. They would undertake recruitment, consent procedures, assist with conveying questions, interpretation/translation of responses, all while following COVID-related precautions about which they received training from the investigator team. The difference in approach across the three field sites (the first pre-COVID-19) and the latter two post-COVID-19 is presented in table 1 along with the sample achieved during fieldwork; and details of the preparation for and conduct of (revised) fieldwork are summarised in table 2.
Comparison of fieldwork approach across three Hamlets before and during COVID-19
Preparation and conduct of data collection for second and third case studies (during COVID-19)
Comparing in person and telephonic fieldwork
In executing our revised fieldwork plan, we had a number of instructive experiences, which are summarised below. In most cases, we would begin by identifying a challenge, but on further reflection, we would see that there were positives associated with the experience as well. We are mindful that there is no clear way to test this impression we have had, as there is no obvious counterfactual for our fieldwork. However, having conducted one case study pre-COVID-19, we used the former as a frame of reference to interpret reflections related to the latter two cases (see table 1).
Adjusting to technology-driven fieldwork posed restrictions but also enhanced the agency and comfort of participants in some ways
For one, given our reliance on phones, our ability to engage with participants was affected by the technology itself. We faced connectivity and noise disturbances due to rain, which was among the reasons why we had to repeatedly reschedule fieldwork in the second and third case study. There were also instances where we had to break the interview across a number of calls, which has been flagged in the literature on research during COVID-19.8 Some of the female participants had to attend other household chores, which meant that each interview took place over multiple sessions. Had we done this in person, we would likely have been accommodated in a single sitting.
At the same time, we were mindful that in case studies two and three, we were unable to build rapport with the participants as we had in the first case study (which involved repeated visits to the community prior to fieldwork). This was particularly problematic in relation to consent taking. For instance, in our first case study, potential participants did raise questions regarding our study that were directed towards us while in the case studies 2 and 3, the process of consenting itself was somewhat rushed, in part because the community researchers had already had preliminary discussions with participants and were known to them. This, however, raised concerns for us on our ability to comprehensively consent participants. We did note, however, is that the main query participants tended to have in our hamlet 1 was in relation to signing consent forms. Participants sought clarification on what the consent form itself was, why they really had to sign it and so on. In our revised COVID-19 protocol, we used recorded oral consent, with which there appeared to be greater comfort regardless of the COVID-19 situation. Participants were mainly interacting with known community researchers: this we feel allowed them to be more candid—there were occasions where they felt more raise sensitive topics like the discrimination faced by tribal persons by ‘outsiders’ (like doctors and health workers). There were also cases where participants refused to answer questions, and stopped interviews. This was much less the case in our face-to-face interviews and/or discussions, where the impression we got—implicit, not explicit—was that participants were concerned that we had come ‘all the way’ to hear from them and they should not let us down.
Study design changes attributable to COVID-19 restrictions were brought about, but also gave us critical insight into the impact of COVID-19 and related outbreaks
By the time our fieldwork was underway, the hamlets for our second and third case studies were adjacent to and very likely to become a containment zone. According to the COVID-19 protocol, group gatherings were only allowed for up to five individuals in containment areas. Having anticipated this in our revised protocol, we no longer sought to have FGDs of 7–10 people and switched our recruitment to group interviews of five individuals. However, when we carried out some practice/mock fieldwork, it became clear that if we were to rely on mobile phones, we would not be able to hear five individuals who were socially distanced. We therefore decided to only have small group (ie, up to four persons) or individual interviews, and reflected this in our revised protocol.
Notwithstanding this, through our community researchers, we were able to get a sense of how COVID-19 was affecting the Kattunayakan community and further, to have conversations with community members in this period about masks, hand washing, social distancing, while also learning about how the pandemic was in fact impacting them. In the second hamlet, the period preceding COVID-19 had seen an outbreak of Kyasanur Forest Disease (KFD) resulting in the death of a hamlet resident. The similarities and differences between KFD and COVID-19 were very much on the minds of a number of our participants and gave us the opportunity to try to understand the perspectives associated with the death from the perspective of community members as well as health system actors. While community members saw the KFD incident as indicative of health system malfunction/dysfunction, health system actors used it to talk about the extents that they went to in order to support the deceased. Apart from this, we were also informed of stoppages and disturbances in continuity of care specifically attributable to COVID-19, which was also a critical finding. Time and logistical constraints (not having our community researcher be able to continue) precluded us from going back to the first hamlet to understand the impact of COVID-19—something we would have liked to do within our research timeframe. It is not clear, moreover, if and how our findings would be different if we could carry out FGDs as well in all of the hamlets.
There was de fact greater reliance on community researchers, which meant we ceded control (to Kattunayakans!)
As aforementioned, we had reposed a great deal of faith in our community researchers—which in turn required ceding a lot of our control over fieldwork. For one, although we had indicated to them to ensure privacy, we were not able to directly oversee this and had to take community researchers in good faith. Moreover, when sensitive topics came up, particularly in relation to individual experiences of health seeking, using a telephonic mode felt impersonal and precluded us from using visual and other cues to calibrate our questioning and responses, as we would have been able to in a face-to-face interaction. So in a sense we were perhaps getting the community researchers’ impressions in hamlets 2 and 3 rather than those of community members.
On the one hand these experiences and contingencies may be seen as aberrations or as weaknesses of our fieldwork. But taking a step back, we felt that particularly for a population of this sort, handing more control to the community researcher—who was a community member could also be interpreted as a positive. We were not, in a sense imposing our interpretations and values on the process. This came out when we compared our fieldwork across hamlets. In the first hamlet (where we were involved with fieldwork), the role of the community researcher was ancillary. However, in the latter two hamlets, our opinions and wishes with respect to the conduct of the interview were also secondary to the dynamics between community members in the roles of researcher and participant(s). This was a somewhat counter-intuitive but powerful realisation for us, particularly since even without COVID-19, our methodological approach did not allow us to embed ourselves for a long period of time with the community (as this was not an ethnography). Even in that situation, perhaps, the power dynamics would be still different from what we experienced.
Indeed, the prospect of communities and community based organisations partnering with research organisations is a larger question that colleagues in our broader research team were already reflecting on—raising the importance of context and the complexities surrounding the decision to engage with researchers in the first instance.12 The anthropological and postcolonial studies literatures make mention of the ‘native informant,’ the ‘insider’ on whom the academic enterprise depends but who as a result faces unique pressures, exerts labour and inhabits vexed roles in relation to those they are informing and those they ‘represent.13 14 ’
In our study, there was noteworthy variation—in hamlet 2, where the community researcher was a young male pursuing higher education and possessing strong motivation to support his community, we received a number of important analytical and logistical insights. For instance, we had detailed commentary from our community researcher in review meetings regarding the socioeconomic conditions of the community, as well as cultural practices, and employment patterns. In the first hamlet, we gained this understanding through a visit (an etic understanding), but through the narrative of the community researcher in hamlet two, we had more of an insider, emic perspective. His observations were continuous and particularly when the context shifted—movement restrictions, illness, etc due to COVID-19—he kept us apprised. The community researcher for hamlet three did not have the same inclination to describe her situation and hamlet in detail: her level of motivation was lower as she was doing this more as a form of income generation. We did not receive additional contextual information about hamlet three from her. The variation in approaches was instructive for us (with respect to the dynamics of gender within the community itself, as well as the possibility that opportunities, aspirations and motivations of young persons in the community may also be gendered). We plan to consider gender dynamics in future fieldwork that we carry out with Kattunayakan and other communities in our health systems research—with or without COVID-19, using or not using telephony.
Conclusion
In this instance, with funder obligations requiring us to complete the study, we carried out an adaptation from in person to telephonic methods of investigation to understand health system access barriers among Kattunayakan communities in northern Kerala. In short, we got it done. But, would we do it again? Perhaps yes. We did not have the opportunity to embed ourselves in long term ethnographic fieldwork with the Kattunayakans (given the continuing scenario with COVID-19), which would have been ideal. However, working with community researchers and letting them guide our fieldwork offered potential in shifting the power hierarchies embedded in public health research. The onus is now on us to sustain, nourish and extend those relationships so that we may all be able to grow our knowledge and contribution to the health system—with and without technology.
Data availability statement
The authors confirm that the data supporting the findings of this study are available within the article.
Ethics statements
Ethics approval
George Institute for Global Health (007/2018), individual informed consent was obtained from all participants.
Acknowledgments
We would like to acknowledge our research team, comprising Ms. Reena PK, Mr Rajeev R, and Ms Jaya R for all their wisdom and support. Thanks are also due to Mr Manoj, Dr Abhilash, Sr Sindhu, Sr Delcy, Ms Prabhavati, Mr Aneesh Sasidharan, Dr Hari Sankar, Mr Jaison Joseph and Dr Tanya Seshadri.
Footnotes
Handling editor Seye Abimbola
Collaborators Reena PK; Rajeev R; Jaya R.
Contributors DN led the project, conceived of and outlined the paper, making substantial edits and final revisions. GB and U were investigators on the project, contributed to the main points outlined in the paper and approved the final version.
Funding This research was funded by a grant from the Indian Council of Medical Research (1/2016/ICMR-ICSSR-SBR).
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.