Agricultural work exposes workers to various physical, chemical, and biological hazards. Soil, as a contaminant reservoir, is understudied, and current risk assessments primarily focus on inhalation and dermal exposures, neglecting incidental soil ingestion. Existing soil ingestion estimates for the general population may not accurately represent the high exposures experienced by agricultural workers. The diversity of agricultural contexts and tasks necessitates a more systematic approach to exposure estimation. This study aimed to address this gap by developing and applying a novel meso-activity-based questionnaire to quantify exposure factors related to soil contact and ingestion among agricultural workers. The researchers sought to improve soil ingestion exposure estimates by incorporating task-specific data into exposure models.

The literature highlights the understudied nature of soil as an exposure pathway in agriculture, particularly regarding incidental ingestion. While research exists on dust exposure, fewer studies characterize the relationship between soil and dust exposure. Inconsistencies in definitions of soil and dust further complicate this issue. Epidemiological studies have mainly focused on inhalation exposures among pesticide applicators, neglecting soil as a contaminant source for ingestion and dermal exposures. Current risk assessment guidelines for agricultural workers, such as those from the Pesticide Handler Exposure Database (PHED) and Agricultural Reentry Task Force (ARTF), primarily address inhalation and dermal routes, omitting incidental soil ingestion. Existing soil ingestion rate estimates for the general population have low confidence, and there is limited empirical data for highly exposed occupational populations like agricultural workers. Default soil ingestion rates used in risk assessments lack validation and are not easily adaptable to the variable tasks involved in agricultural work. Previous studies have explored matrix approaches for pesticide exposure by task and crop type, but this approach has not been widely applied to soil contaminant exposures. A previous qualitative study identified factors influencing soil exposure, providing a framework that this study aimed to quantify.

This study used a meso-activity-based, season-specific soil contact activity questionnaire administered to 38 fruit and vegetable growers in Maryland. Recruitment was conducted through Maryland's Best website and previous study participants. The questionnaire, administered via telephone due to COVID-19 restrictions, collected data on frequency and duration of six meso-activities (bed preparation, planting seeds, transplanting, irrigation, weeding, harvesting), soil contact intensity, glove use, handwashing practices, and non-dietary ingestion events (soil in mouth/face, sampling produce, eating onsite). Data were collected and managed using REDCap. Kruskal-Wallis tests, Shapiro-Wilk tests, Pearson correlations, and Kendall's tau tests were used for data analysis. Exposure assessment involved calculating average daily doses (ADDs) using three methods: 1) daily soil ingestion rate, 2) hourly soil ingestion rate (differentiating indoor/outdoor time), and 3) hourly-task-specific ingestion rate (adjusting for soil contact intensity). Empirical exposure estimates were complemented by Monte Carlo simulations using data from the NHANES survey.

Thirty-eight growers participated, with varying demographics and farm sizes. Growers reported the greatest number of hours worked during summer and least during winter. Irrigation was reported as the most frequent meso-activity, while bed preparation had the longest duration. Transplanting and weeding involved the greatest soil contact, possibly due to kneeling postures. Glove use and handwashing were common, but varied among activities. Soil ingestion was more frequently reported during summer and spring. Most growers reported ingesting less than 10 mg/day, but some reported higher amounts. Statistically significant differences in exposure factors were observed across seasons. Total annual exposures estimated by the three methods were similar in order of magnitude, with method 3 (hourly-task-specific) generally yielding the lowest values. Method 3 revealed high variability in task-specific ADDs across seasons, with no clear patterns of which activities contributed most to exposure. Monte Carlo simulations produced ADD distributions across seasons, methods, and activities.

The study's findings emphasize the significant variability in soil exposure among agricultural workers due to differences in both the frequency and intensity of tasks. The meso-activity-based approach provides a more accurate representation of exposure than using a single daily ingestion rate. The three methods used for estimating ADDs yielded comparable results, suggesting that while task-specific hourly rates are more detailed, simpler methods could also offer useful estimates for risk assessment. The study highlights the importance of considering both task frequency and soil contact intensity when assessing exposures. The observed variability emphasizes the need for task-specific interventions. The open-ended questions in the questionnaire provided valuable qualitative insights about growers' perception of soil versus dust, handwashing practices, and the impact of multitasking.

This study successfully generated meso-activity-specific exposure factors for soil exposure among agricultural workers and demonstrated their application in exposure modeling. The results highlight the importance of considering task-specific variability in soil exposure assessment. Future research should focus on validating the task-specific ingestion rates, characterizing subconscious soil exposure factors, and using time-activity diaries to improve exposure estimates. The study also emphasizes the need for further research on the influence of other factors such as farm size, tools used, and growing practices on soil exposure.

The study relied on growers' self-reported data, potentially introducing recall bias. The limited sample size and geographical focus might limit the generalizability of the findings. The questionnaire only covered six meso-activities and did not capture all possible activities related to soil exposure. The study could not differentiate between intentional and unintentional soil ingestion. The use of a hypothetical contaminant concentration limits the direct applicability of the ADD estimates to specific contaminants. The study's reliance on grower's self reported soil ingestion values may have introduced bias and may not reflect the actual soil ingestion habits of the farmers studied.