Key considerations for assessing soil ingestion exposures among agricultural workers

The study addresses the underrecognized route of contaminant exposure via soil ingestion among agricultural workers, who frequently and directly contact soil during tasks such as cultivation and maintenance. While exposure to contaminated soils can occur through dermal, inhalation, and ingestion pathways, ingestion is hypothesized to be dominant. Traditional dietary assessment tools are inadequate for incidental soil ingestion, and existing exposure science relies on time-activity factors and non-dietary ingestion behaviors—areas with limited adult data and heavy reliance on children’s behavior patterns. Little is known about adult growers’ non-dietary ingestion patterns, and task-specific factors in agriculture are undercharacterized. This study aims to qualitatively identify key tasks, behaviors, and contextual factors that shape soil contact and ingestion among growers to inform improvements to exposure estimation tools.

Prior work and EPA guidance distinguish soil, outdoor settled dust, and indoor settled dust for exposure estimation, with soil ingestion encompassing soil and outdoor dust and dust ingestion referring to indoor settled dust. Research has focused on general population activity patterns and children’s hand-to-mouth behaviors, with limited studies on adult non-dietary ingestion in occupational settings. The translation of growers’ terminology (soil, dirt, dust) to exposure science definitions is unclear and may risk misclassification. Limited observational studies in industrial and agricultural settings suggest PPE and attire influence hand and object-to-mouth behaviors that can transfer soil to the mouth. There is a gap in characterizing agricultural tasks and their influence on soil ingestion.

Qualitative in-depth interviews (IDIs) were conducted with fruit and vegetable growers in Maryland (January–February 2020). Purposive sampling used the Maryland’s Best database and an internal database from prior Baltimore City research to identify eligible participants. Recruitment occurred via email and networking at conferences and community events. Inclusion criteria: ≥18 years; current farm owner/manager, employee, or community gardener in Maryland; completion of food production tasks (e.g., planting, harvesting, weeding, mulching) within the past 12 months; expectation of ongoing engagement in farm activities in the next 12 months. Informed consent was obtained; interviews were audio-recorded. Participants completed a brief demographic questionnaire and received a $20 gift card. IRB approval: Johns Hopkins IRB (IRB00009866). Interviews were conducted onsite by the first author using a semi-structured guide covering farm operations, labor distribution, typical workdays, detailed descriptions of tasks (planting, irrigation, weeding, harvesting), soil contact (including incidental ingestion), methods of modifying contact (PPE, attire, hand hygiene), and health/safety concerns. Initially, questions used the term “soil”; through iterative transcript review and reflexive journaling, probes were added after the first two interviews to explore participants’ use of “dirt” and “dust.” Audio files were transcribed via NVivo transcription and verified by listening against recordings; one audio file was lost prior to transcription and excluded from analysis. An adapted framework approach guided analysis. The first author applied deductive codes based on interview guide topics, then inductively identified emergent themes. Six frequently mentioned tasks (bed preparation, planting, pest management, harvesting, weeding, produce handling) emerged. Codes were aggregated, and a second round of inductive coding identified themes within aggregates. Analytical frameworks were developed by mapping a priori exposure concepts (definitions of soil/dust, routes of exposure, hierarchy of controls) and emergent themes, with co-author review and refinement.

• Sixteen IDIs were completed (duration 21–92 minutes; mean 50 minutes). Participants were predominantly female (n=9), full-time workers (≥35 h/week; n=10), and urban (Baltimore City) growers (n=9). Age distribution among those reporting: 20–29 (n=4), 30–39 (n=4), 40–49 (n=1), 50–59 (n=2), 60+ (n=4); one declined to report age. Education was largely bachelor’s degree (n=11). - Four emergent themes: (1) Variability in growers’ descriptions of soil and dust: Growers often used “soil,” “dirt,” and “dust” interchangeably, with meanings not consistently aligning with EPA Exposure Factors Handbook definitions. This presents translational challenges and potential exposure misclassification (e.g., some used “dirt” to describe indoor dust or outdoor deposited materials; “soil” connoted living, managed media; “dust” had negative connotations like irritation or hazards such as lead). (2) Variability in growers’ soil contact: Soil ingestion occurs within multi-route exposure, with frequent descriptions of dermal and inhalation contact. PPE and attire likely influence hand/object-to-mouth behaviors. (3) Growers’ concerns regarding soil contact: Concerns included irritation and hazardous dusts (e.g., lead), and distinctions between positive perceptions of “soil” versus negative perceptions of “dirt” and “dust.” (4) Growers’ practices to modify soil contact: Strategies included PPE use, attire choices, and hygiene practices to reduce contact and potential ingestion. - Task-specific insights: Six routine tasks were identified—bed preparation, planting, pest management, irrigation, harvesting, and produce handling. Urban growers most often cited planting as the most soil-contact-intensive task (all but one; the other cited bed preparation). Rural growers most often cited bed preparation and pest management/weeding as highest contact. Environmental and behavioral factors within tasks (e.g., climatic conditions; barefoot walking, hand-to-mouth frequency) may modify ingestion rates. - Conceptual framework: Soil contact events were classified by worker intention (intentional vs. unintentional) and cause (environmental vs. behavioral), then mapped to exposure routes. Intentional ingestion was rare (one participant occasionally tasting soil); no pica was reported. Behavioral factors commonly facilitated transfer to the mouth; environmental factors dominated inhalation-related contact. Findings suggest that soil ingestion assessments should explicitly query other exposure routes that co-occur and influence ingestion rates.

One audio recording was lost prior to transcription and excluded from analysis. The interview probe to differentiate “soil,” “dirt,” and “dust” was added after the first two interviews; although deemed unlikely to differentially impact content, this change could introduce minor variability across interviews.