Introduction/Objective Farming is a highly variable occupation, with many tasks and exposures, making exposure assessment for epidemiologic studies challenging. We developed and deployed a smartphone app to collect real-time information on intermittent agricultural activities to characterize farming task variability.

Methods We recruited 19 male Iowa farmers, age 50–60 years, to log their farming activities in the app on 24 randomly selected days over 6 months. We populated the app with 350 farming activities; 152 activities were also linked to contextual questions (e.g., pesticide application method, PPE use). We calculated descriptive statistics on the number of activities reported and their duration.

Results The farmers provided activity information for 283 days. The farmers submitted 1,331 activities, representing 124 unique farming tasks. The median duration of a logged day was 545 minutes (interquartile range, IQR: 431–698). The median number of tasks reported per farmer was 18 (IQR: 5–31), with a median of 4 activities per day (maximum 17). The median duration of activities was 63 minutes (IQR: 32–133). The three most frequently reported tasks were related to animal work (36% of activities), transportation (12%), and crops (10%). The tasks with the longest daily duration were planting crops (median: 415 minutes), mixing/loading/applying pesticides (365 minutes), and loading corn (270 minutes). The shortest tasks (median duration 10 minutes), were fueling trucks, collecting/storing eggs, and tree work. Over 36% of the submitted activities also included responses to contextual questions; these were most frequently about feeding animals (56%), transportation (25%), and mixing/loading/applying pesticides (6%).

Conclusion Our findings show that it is possible to collect real-time, intermittent activity information over the span of several months. We captured most of the farming day and, as expected, observed substantial heterogeneity in activities and their durations, highlighting the need for individual-level activity data when evaluating risks in farmers.