Successful completion of the cross-sectional nutrition/microbiome study in Buli, Zhemgang!

We are delighted to convey that we have successfully concluded the cross-sectional nutrition/microbiome study in Buli and nearby surroundings where we sampled and interviewed 223 people within the course of one month. Success can be attributed to the fact that the two main researchers in the field have been living within the community for nearly a year before starting this particular part of the study. In this way, our PhD-student Kanchan Kattel and her research collaborator Tshering Tashi have acquired a deeper understanding of local sensibilities, activities, and the context as well as build relations of trust, all of which will generate better data and a culture-sensitive understanding and interpretation of these data. To achieve such an integrated understanding of the entangled microbiosocial nature of food, nutrition and microbiomes within their socio-ecological environment is precisely the main aim of the EATWELL project. Yet, again, achieving this second important milestone has not been a straightforward matter, claiming many logistical, administrative, and scientific decisions as reality struck back in several ways.

After the successful completion of the cross-sectional nutrition/microbiome study in Laya, there was hardly any time to rest. Many bills had to be submitted and resubmitted for refund, the gathered data had to be cleaned, meetings with the governors of 3 Dzongkhags were held to prepare for the cross-sectional study in the urban food environments, and of course, the logistics for the cross-sectional sample and data collection in Buli, Zhemgang, had to be prepared, all this within the course of less than two weeks.

In addition, we had to process and wait for the paperwork and permits for the continuation of the longitudinal part in Laya as our participants listed different locations than what we had processed earlier. To follow them and sample and interview them in these locations, we needed to request an updated support letter from the Governor of the district to be able to apply for the army permit, permit from the Department of Forestry, and the Immigration to visit our participants where they are at the given time points.

Hence, as the team waited for these permits, we decided to include our Research Assistants (RAs) in the work in Buli, in the hope of minimizing the many delays incurred during this project. This meant that we had to rearrange the transport, food supply and place of stay in Buli at short notice. The sample kits and the materials were already sent early March when Kanchan and Tshering left to Buli.

In March, when two members of our team, Kanchan and Tshering, arrived in Buli, we noticed that our plan to store the freezers and generators in the local hospital would not be feasible. We had to quickly look for an alternative space and eventually Kanchan and Tshering found a disused waste room in the Gewog office. As the space was far from ready to use as a storage room for the -80 freezer, it required considerable work to make it suitable for this new purpose.

Renovating the space meant tackling several practical issues: installing a new electricity meter, ensuring a stable power connection, and covering up the metal fence to protect the place against rainstorms. Tshering took the lead on re-constructing the sample room in close conversation with the PI, Wim Van Daele. He identified a local person to work with to arrange the electricity installation, and he filled the freezers with saltwater bottles to stabilize the freezer’s internal temperatures: Tshering also arranged for someone from nearby home to monitor the room in case of power outages, ensuring the generator could be turned on promptly when needed.

While Tshering worked on the logistics, Kanchan focused on using the survey data from Laya to calculate standard recipes. Firstly, she identified commonly reported/consumed recipes that are prepared similarly and could also be used in Buli. Then, through highlighting the data points for similar recipes often coming in many different names, flagging errors, and identifying potential outliers, classifying base and supplementary ingredients, she calculated an average proportion of ingredients for devising standard recipes. Recipe calculation proved to be time-intensive and required both technical and contextual judgment. Choosing the right variation of a recipe, identifying RA’s errors, and distinguishing outliers all demanded a deep familiarity with the food itself. In case of doubt, Kanchan and Tshering verified average proportions visually by asking host families whether the calculated amounts of ingredients like oil or cheese seemed realistic. The final set of eleven standard recipes was uploaded to the input database on globalfoodmattersdatabase.org. This database, which brings together food composition data, recipe details, portion conversion factors, and tagging systems, is linked to the 24-hour recall survey tool, INDDEX24. Once the updates were made, the tablet interface used by enumerators displayed these standard recipe names in a drop-down menu, saving both research assistants and respondents’ time and effort that would otherwise be spent describing ingredients and their quantities in detail during the survey.

Although Kanchan had been working with the database for over a year, the urgency to prepare it for the Buli survey brought an added intensity to this phase. The process of adding new recipes, food descriptors and portion conversion factors involved close communication with the technical team of FHI360 at Intake for troubleshooting and for assistance in these complex technical procedures.

At the same time, Tshering continued measuring actual food items available in the field and identifying proxies to complete the portion conversion factor table. Kanchan provided real-time supervision to this work and double-checked calculations as they came in. They also expanded the database by adding seasonal foods available in the spring and early summer—an important step for accurate food matching during the upcoming Buli survey. This groundwork continued until mid-May, when the larger team of RAs departed from Thimphu and made their way to Zhemgang for a long-waited phase of work.

Before travelling to Buli, the team in Buli still had to buy food supplies, refill the gas cylinders, pack additional kits, and adjust the booking of the bus to now transport 10 RAs and of the vehicle to carry the luggage and a regular freezer to Buli. On Sunday 18 May, we packed the Bolero pick-up truck with the regular freezer brought from Laya, gas cylinders, luggage, and research materials. The following morning, Wim picked up the RAs with the bus, loading more luggage onto the pick-up truck and the bus.

The team left at 7.30 am and had some breakfast in Lobesa and shopped for additional fresh food at the Wangdue vegetable market to bring to Buli, as only onions, potatoes, and chilies were widely available there around this period. Even though Buli has a weekly supply of fresh fruit and vegetables from a food truck, it would probably not suffice to supply extra food for the entire team.

After a long journey, we all arrived in Zhemgang town where the travelling team was joined by Kanchan and Tshering. We stayed in two locations, including a hotel above a karaoke bar with people singing until 3 am, sometimes nice and sometimes quite false. After a short night, we had a courtesy call with the Dasho Dzongrab, representing the Dasho Dzongdag or governor of the Zhemgang district, and other district officials. We explained the status of the ongoing work in Buli as well as the further plans and sought further support of the district administration. After the successful meeting, we visited the beautiful old part of the town, and after lunch we left for Buli driving on a bumpy small road for 3 hours.

After settling down in the homestay we started preparing the work: organizing the community meeting to recruit 240 participants for the cross-sectional part in Buli and discussing the coordination of the work with Dr Neyzang and Kanchan. Since we had some work correcting codes after the RA’s had confused them in Laya, we also discussed another way of proceeding in assigning open participant codes to the RA’s who would then assign these individually to the participants. The number of mix-ups in the codes remained the same, however, these were sorted out as well.

Two days later we had the community meeting where we recorded about 80 potential participants and as soon as all were recorded, the RAs were assigned ‘their’ respondents. They immediately started calling them to make appointments for the informed consent meeting and distribute the sample kits. Soon thereafter, the train was rolling, and samples got collected and surveys conducted.

We expected the situation in Buli to be less complicated than Laya, but we were sure that other logistical challenges would be there, and this proved to be the case. Buli lies in Zhemgang in the southeast of Bhutan, which is a subtropical area and a 2-day journey away from Thimphu. It is much warmer than in Laya and therefore the cold chain has been more difficult to maintain. Moreover, the power supply is often interrupted, even though mostly briefly, but this has required more intense following-up of the freezer and using the generator. We therefore had to again adjust the electricity supply and modify the operation of the sample collection.

Given the heat during the days, it was not sure that the regular freezer could keep the samples sufficiently cold during the daytime, hence we used the other -80 freezer brought from Laya as the freezer into which the RA’s could deposit the cooling elements and the samples. Opening this freezer regularly meant that it could only maintain a temperature between -30 and -40 degrees. This was anyhow better than using the regular freezer which could keep the samples at -23 degrees at best. The samples were anyhow transferred to the main -80 freezer at regular time points. Whichever freezer used, the use of two freezers meant that we had to install an additional electricity line and socket for the second freezer. Dividing the power supply of 32 Ampère over two sockets meant that it decreased to 16 Ampère each and that was not sufficiently powerful to keep the -80 freezers at the required temperature as per our protocol. Hence, we had to increase the main power supply to 63 Ampère. In addition, we had to upgrade the extension cord for the times when the generator was used, which was quite often.

Given the heat in the storage room, which was securely packed in plastic to protect the materials against rainstorms, we also had to make a compromise and lower the plastic protection for a better ventilation of the storage room so that the temperature could be better maintained.

Even though Buli is a concentrated settlement in the middle of the large forest cover of the Zhemgang district, we were increasingly becoming aware during the recruitment that the constitution of the village community was perhaps more diverse than in Laya. In Buli, many intermarry with people from other locations in Eastern and Southern Bhutan, whereas in Laya people predominantly intermarry between the different Chiwogs or villages of Laya. Hence, there is a greater diversity in regional backgrounds within the population of Buli, which thereby includes more people with diverging food habits and thus may have a gut microbiology that evolved in relation to a different socio-ecological environment prior to their migration to Buli. To avoid introducing too much diversity in our already limited sample size of 250 people, we decided then to include migrants from these other areas, setting the criteria that they have already lived in Buli for at least 3 years continuously, hypothesizing that their gut ecology and eating habits would have sufficiently adjusted to the socio-ecological environment of Buli and no longer diverge too much from the life-time inhabitants of Buli and its near surroundings. Hence, the key assumption that dwelling and food habits would have adapted over the course of three years to the rest of the Buli community became the premise for introducing the people with different backgrounds into this part of the study, yet while carefully recording their migratory history. This would allow us to analyze nutrient intake and gut microbiomes at sufficient statistical levels and aggregate or disaggregate different subgroups should we find that the food habits or gut ecologies would differ considerably, even after 3 years of dwelling and eating in Buli. More on this will follow in our publications during and after our analyses.

The diversity in backgrounds and Buli also raised the practical issue of the different languages spoken among the community members of Buli, which already required us to hire Khengkha speaking RAs as this is the main language spoken in this area. Yet, within this language there are also many dialects which can be hard to understand, which was the case with one of our RAs who spoke a Khengkha dialect from nearby southern Trongsa. Besides that, the people who have migrated to Buli were perhaps more comfortable with Tshangla (the main language for Sarchops from the East) or with Nepali (for Lhotshampas or ethnic Nepali from the South), but then we conducted these in Dzongkha, since it took considerable time to train the Dzongkha speaking RAs in correctly phrasing the survey questions and probing, and this learning process had to be repeated in Khengkha for the Khengkha speaking RAs, and so we could not spend much more time on training them in conducting the survey in these two other languages as well. Hence, we could not cater to all the local languages spoken, but we did so for the two main spoken ones.

As we had the meetings for informed consent, we also gradually recruited more participants and thus expanded the sample size. Again, when they received a call from ‘their’ respondent, sometimes as early as 5 am, that they had collected their stool sample, the RAs had to get up and collect the kit and conduct the 24-hours recall and the integrated nutrition/microbiome questionnaire. They had long days and had to go in the scorching sun or in the pouring rain.

Yet, when we had collected around 150 consent forms and as the sample and data collection was at full speed, we started to feel some saturation, meaning that we did not find many new potential participants and the recruitment started to slow down. We then decided to withdraw the longitudinal RA’s from Buli and bring them back to Thimphu for a rest, before leaving for Laya for the continuation of the longitudinal study there. Kanchan Kattel joined them to return soon thereafter to Norway as per the Norwegian regulations. Simultaneously, we could combine this with the delivery of the third batch of samples to avoid the cost of a separate delivery, which would have included expenses for the transport, lodging and fooding of the driver and the RA managing the cold chain until it reached the Royal Centers for Disease Control in Thimphu.

The saturation in recruitment led to discussions on how we could further expand our potential participant base to reach the numbers required for a detailed subgroup analysis in terms of nutritional intake and the gut microbiome. Several of the possibilities—students above 18 years at the boarding school and monks as a target group at the temple—were not possible to include since that would have required additional permits and approvals, so we started exploring the possibility for recruitment at nearby villages or hamlets, something which we had considered earlier as possible back-up options. Yet, our ethnographic observations showed that the main nearby village, Kikhar, was quite different from Buli in terms of the food system, related to the lower altitude of the place. Hence, including this community would have introduced a more radical difference in dwelling environment and food habits, which might complicate the integrated nutrition/microbiome analyses, compounded by a lack of solid ethnographic data to make sense of the data. Moreover, the logistics of conducting the sample collections and surveys there was very complicated and even though we could have done an aggregate analysis with Kikhar we thus decided not to include it.

Yet, we expanded to some hamlets nearby and two other villages which from our few observations seemed more similar and this was confirmed by Ben Hewitt, a PhD student in Linguistics at the University of Sydney studying the Khengkha language in the area. Hence, during the final week, six RAs went to Buli Goenpa and Tali, where people are closely related to the community of Buli and where the dwelling environment and food system is more similar than in Kikhar. This additional recruitment added around 30 people to our number of participants. This part was intense, as some got up at 3 AM to prepare breakfast and lunch packets to take to these locations, which took up to 2-hour drives back and forth to deliver the samples in time in Buli. Despite the challenges, we successfully completed 223 surveys in Buli, Zhemgang, bringing us much closer to our target!

What this narration of decision-making processes and logistical challenges shows, is that science in practive may not be as straightforward as it may seem in the scientific publications, omitting the multiple frictional factors that come into play in the generation of published facts. Yet, some fields of social sciences of science highlight these nuanced processes of decision making critical in the generation of facts and knowledge production. Our radical interdisciplinarity in this project explicitly considers the situated logistical and practical details which shape these decision-making processes which are understood through earlier ethnographic fieldwork to make well-informed and precise decisions, thereby describing how knowledge/practices are established when realities strike back at different aspects. Indeed, a key premise in EATWELL is that radical interdisciplinarity is a precondition to better (social, nutritional, and microbiological) sciences.

On 19 June, two cars collected the last batch in a fully filled -80 freezer and the remaining four RAs to bring to Thimphu and this signaled thus the end of our cross-sectional part in Buli. Yet, many challenges arose on the way back due to the rains which caused some smaller landslides on the way and where the RA’s joined in cutting the fallen trees and clearing fallen boulders so that they could proceed. Happily, everyone reached home safely to enjoy the successful completion of this second milestone in the EATWELL nutrition/microbiome phase, despite the numerous challenges.