**4. Discussion**

With the call and opportunity to promote inclusivity in microbiome research, an appropriate, low-cost and appropriate stool collection method is warranted for use in rural and low-resource settings [8,11]. The collection of stool specimens from a large cohort of rural-dwelling women could present challenges regarding participant recruitment, specimen collection, retention, and managemen<sup>t</sup> of staff resources. We describe the development, assembly and use of a simple, low-cost at-home stool collection kit for rural or low-resource settings where modern toilets with seats are unavailable. Using this at-home stool collection kit was reported as easy and safe.

Other authors have reported using at-home adult stool collection tools but are also limited by the availability of a modern toilet with a toilet seat [5,6]. Further, contemporary over-the-toilet seat stool collection supplies (e.g., flushable collection paper secured to the toilet seat) cannot be procured in some countries, such as Cambodia, and instructions are not available in different languages. In rural and low-resource settings, even when improved sanitation infrastructure is built, there is still a lack of facilities allowing for straightforward and sterile stool collection. Our kit was <\$5 USD, thus, it is a low-cost option for single and multiple follow-up stool specimen collections.

To our knowledge, no authors have reported on the development and use of a simple adult stool collection kit for use at an individual's home in a rural or low-resource setting. There is no consensus or guidance on an appropriate or detailed method of stool collection in rural or low-resource settings. In other reports in rural and low-resource settings, the general practice is to collect the stool specimen at the local health centre [12], which may be unfeasible for studies with large sample sizes and for women who cannot defecate 'on demand'. Our stool collection method is novel, as it allows participants to independently collect a stool specimen in the comfort and privacy of their home and when they feel the 'urge' to defecate. This approach also reduces the burden on local health facilities and research staff and is optimal for use in large-scale research.

Ensuring that specimen collection methods are culturally acceptable is essential to improve participation and minimize attrition rates in the study population. On account of our high study retention rate (92%) and stool specimen collection rate at 12 weeks (90%), we infer participants generally accepted this stool specimen collection method. However, these findings may have been affected by response bias. We should reiterate that our high study retention rate was likely due to our experienced field research staff's strong rapport with study participants. Detailed staff training resulted in the clear communication of stool collection instructions. We also recognize the limitations of this method of stool specimen collection. Although the metal pots were stored together in heavy-duty plastic bags to prevent contamination by dirt and debris, they were not stored in a sanitized environment, allowing for possible contamination during storage and transportation. As a lesson learned, we recommend that collection pots/containers be wrapped in protective sealing and stored in clean areas. Alternatively, the collection pots/containers could be sanitized prior to defecation at each specimen collection time point, if participants were provided with such materials. Secondly, our research group provided clear bags for the transportation of stool specimens to the health centre. Still, most participants opted to put the clear participant labelled bag inside in their own small black garbage bag for privacy. Therefore, we recommend supplying a discrete, non-opaque bag or container for

participants to return specimens to the study staff. Lastly, it would be advantageous to conduct a standardized assessment of user acceptability of this stool specimen collection technique in future work.
