Background Data reported in the published literature have frequently been used qualitatively to aid exposure assessment activities in epidemiologic studies. Analysing these data in statistical models presents statistical challenges because these data are usually reported as summary statistics. Most previous analyses using published data have weighted the summary statistics by the number of measurements, but this does not account for the measurements’ variability. We describe the application of mixed-effects meta-regression models to evaluate the relative contributions of three exposure pathways (agricultural drift, para-occupational, residential use) on published pesticide concentrations in the house dust of homes in agricultural areas.

Methods We abstracted pesticide house dust concentrations reported as summary statistics (e.g., geometric means (GM)) from studies in North American agricultural areas published from 1995–2015. We analysed these data using mixed-effects meta-regression models that weighted each summary statistic by its variance. The dependent variable was either the log-transformed GM (drift) or the log-transformed ratio of GMs from two groups (para-occupational, residential use).

Results For the drift pathway, the predicted GM decreased 35% (95% Confidence Interval [CI]: 19–48; based on 52 statistics from 7 studies) for each natural log(ft) between homes and fields. For the para-occupational pathway, GMs were 2.3 times higher (95% CI: 1.5–3.3; 15 statistics, 5 studies) in homes of farmers who applied pesticides more versus less recently or frequently. For the residential use pathway, GMs were 1.0 (95% CI: 0.8–1.2), 1.3 (95% CI: 1.1–1.4), and 1.5 (95% CI: 1.2–1.9) times higher in treated versus untreated homes, when the probability that a pesticide was used for the pest treatment was 0%, 1–19%, and ≥20%, respectively (88 statistics, 5 studies).

Conclusion These findings quantify relative contributions of three pathways in agricultural populations that may be useful for developing pesticide exposure metrics for epidemiologic studies. The meta-regression models can be updated when additional data become available.