UC Berkeley

In sub-Saharan Africa, many families travel to collect water and store it in their homes for daily use, presenting an opportunity for the introduction of fecal contamination. One stored and one source water sample were each collected from 45 households in rural Kenya. All 90 samples were analyzed for fecal indicator bacteria (E. coli and enterococci) and species-specific contamination using molecular microbial source tracking assays. Human (HF183), avian (GFD), and ruminant (BacR) contamination were detected in 52, two, and four samples, respectively. Stored water samples had elevated enterococci concentrations (p < 0.01, Wilcoxon matched pairs test) and more frequent BacR detection (89% versus 27%, p < 0.01, McNemar's exact test) relative to source water samples. fsQCA (fuzzy set qualitative comparative analysis) was conducted on the subset of households with no source water BacR contamination to highlight combinations of factors associated with the introduction of BacR contamination to stored water supplies. Three combinations were identified: (i) ruminants in the compound, safe water extraction methods, and long storage time, (ii) ruminants, unsafe water extraction methods, and no soap at the household handwashing station, and (iii) long storage time and no soap. This suggests that multiple pathways contribute to the transmission of ruminant fecal contamination in this context, which would have been missed if data were analyzed using standard regression techniques.