RESUMO
OBJECTIVES: The objectives of this work were to evaluate (i) the prevalence and frequency of caregiver-reported soil ingestion by children, (ii) whether household flooring material in the bedroom (earth vs. concrete) affected caregiver-reported soil ingestion, (iii) whether caregiver-reported soil ingestion was associated with caregiver-reported diarrhoea and (iv) caregivers' perceptions of their children ingesting soil. METHODS: We conducted 309 household surveys in northern Ghana, including 529 children under five (249 children aged 6-36 months), and measured faecal contamination in soil from 31 households. RESULTS: Among all children, 15% were reported to have directly ingested soil in the past week, including 28% of children aged 6-36 months. Among children reported to have ingested soil, the median frequency was 14 times in the past week, and the median amount of soil ingested each time was half a handful. There was no association between household floor material and whether the caregiver observed a child directly ingesting soil. After adjusting for household floor material and other potential confounding variables, caregiver-reported soil ingestion was associated with caregiver-reported diarrhoea for children under five [adjusted odds ratio (adj. OR) = 3.13, 95% confidence interval (CI) 2.76-3.55] and children aged 6-36 months (adj. OR = 2.61, 95% CI 2.01-3.39). Approximately 83% of caregivers whose children ingested soil reported they thought it was unsafe and were more likely to report stopping their child from ingesting soil, but these responses did not affect the quantity of soil ingested. CONCLUSIONS: Our results suggest direct soil ingestion is associated with diarrhoea independent of household floor material, and separate interventions may be necessary to prevent exploratory soil ingestion.