Using Machine Learning and Nationwide Population‐Based Data to Unravel Predictors of Treated Depression in Farmers DOI Creative Commons
Pascal Petit, Vincent Bonneterre, Nicolas Vuillerme

и другие.

Mental Illness, Год журнала: 2025, Номер 2025(1)

Опубликована: Янв. 1, 2025

Farmers are exposed to numerous stressors that can negatively impact their mental health, leading conditions such as depression. However, most studies examining depression risk in farmers limited by small sample sizes, narrow geographic coverage, and a focus predominantly on male general agricultural contexts. To complement these traditional studies, big data machine learning (ML) advantageously be harnessed. While ML algorithms have shown high accuracy identifying predictors health research, no study has yet applied farmers. We aimed identify key of among the entire French farmer workforce across professional categories, activities, sexes using (XGBoost). A secondary analysis large‐scale administrative databases (TRACTOR project) was conducted. Potential ( n = 128 for farm managers 123 farmworkers) included broad range sociodemographic, lifestyle, occupational variables. The predictor’s importance determined Shapley’s additive explanation. There were 83,592 cases 1,088,561 149,285 5,831,302 farmworkers. Models performed well, with F 1 scores ranging from 0.65 0.94. noted differences, even though several common populations, and/or sexes. top working year, age, sex, experience, job security, income, preexisting conditions. which reflects cumulative external factors (e.g., harsh weather) farmers’ emerged important predictor. These findings highlight potential real‐world modifiable predictors, thus enhancing early detection prevention strategies. By differentiating farming groups, our results suggest tailored interventions could developed better address unique needs various populations. insights inform development clinical tools calculators) assist decision‐making.

Язык: Английский

Investigating Parkinson’s disease risk across farming activities using data mining and large-scale administrative health data DOI Creative Commons
Pascal Petit, François Berger, Vincent Bonneterre

и другие.

npj Parkinson s Disease, Год журнала: 2025, Номер 11(1)

Опубликована: Янв. 8, 2025

Abstract The risk of Parkinson’s disease (PD) associated with farming has received considerable attention, in particular for pesticide exposure. However, data on PD specific activities is lacking. We aimed to explore whether exhibited a higher than others among the entire French farm manager (FM) population. A secondary analysis real-world administrative insurance claim and electronic health/medical records (TRACTOR project) was conducted estimate 26 using mining. cases were identified through chronic declarations antiparkinsonian drug claims. There 8845 1,088,561 FMs. highest-risk group included FMs engaged pig farming, cattle truck fruit arboriculture, crop mean hazard ratios (HRs) ranging from 1.22 1.67. lowest-risk all involving horses small animals, as well gardening, landscaping reforestation companies (mean HRs: 0.48–0.81). Our findings represent preliminary work that suggests potential involvement occupational factors related onset development. Future research focusing farmers high-risk will allow uncover by better characterizing exposome, which could improve surveillance farmers.

Язык: Английский

Процитировано

2
Global research trends on the human exposome: a bibliometric analysis (2005–2024) DOI Creative Commons
Pascal Petit, Nicolas Vuillerme

Environmental Science and Pollution Research, Год журнала: 2025, Номер unknown

Опубликована: Март 8, 2025

Abstract Exposome represents one of the most pressing issues in environmental science research field. However, a comprehensive summary worldwide human exposome is lacking. We aimed to explore bibliometric characteristics scientific publications on exposome. A analysis from 2005 December 2024 was conducted using Web Science accordance with PRISMA guidelines. Trends/hotspots were investigated keyword frequency, co-occurrence, and thematic map. Sex disparities terms citations examined. From 2024, 931 published 363 journals written by 4529 authors 72 countries. The number tripled during last 5 years. Publications females (51% as first 34% authors) cited fewer times (13,674) than males (22,361). Human studies mainly focused air pollution, metabolomics, chemicals (e.g., per- polyfluoroalkyl substances (PFAS), endocrine-disrupting chemicals, pesticides), early-life exposure, biomarkers, microbiome, omics, cancer, reproductive disorders. Social built environment factors, occupational multi-exposure, digital exposure screen use), climate change, late-life received less attention. Our results uncovered high-impact countries, institutions, journals, references, authors, key trends/hotspots. use technologies sensors, wearables) data artificial intelligence) has blossomed overcome challenges could provide valuable knowledge toward precision prevention. risk scores represent promising avenue.

Язык: Английский

Процитировано

0
Using Machine Learning and Nationwide Population‐Based Data to Unravel Predictors of Treated Depression in Farmers DOI Creative Commons
Pascal Petit, Vincent Bonneterre, Nicolas Vuillerme

и другие.

Mental Illness, Год журнала: 2025, Номер 2025(1)

Опубликована: Янв. 1, 2025

Farmers are exposed to numerous stressors that can negatively impact their mental health, leading conditions such as depression. However, most studies examining depression risk in farmers limited by small sample sizes, narrow geographic coverage, and a focus predominantly on male general agricultural contexts. To complement these traditional studies, big data machine learning (ML) advantageously be harnessed. While ML algorithms have shown high accuracy identifying predictors health research, no study has yet applied farmers. We aimed identify key of among the entire French farmer workforce across professional categories, activities, sexes using (XGBoost). A secondary analysis large‐scale administrative databases (TRACTOR project) was conducted. Potential ( n = 128 for farm managers 123 farmworkers) included broad range sociodemographic, lifestyle, occupational variables. The predictor’s importance determined Shapley’s additive explanation. There were 83,592 cases 1,088,561 149,285 5,831,302 farmworkers. Models performed well, with F 1 scores ranging from 0.65 0.94. noted differences, even though several common populations, and/or sexes. top working year, age, sex, experience, job security, income, preexisting conditions. which reflects cumulative external factors (e.g., harsh weather) farmers’ emerged important predictor. These findings highlight potential real‐world modifiable predictors, thus enhancing early detection prevention strategies. By differentiating farming groups, our results suggest tailored interventions could developed better address unique needs various populations. insights inform development clinical tools calculators) assist decision‐making.

Язык: Английский

Процитировано

0