@univ-grenoble-alpes.fr
Laboratoire AGEIS - Université Grenoble Alpes
Exposome, health sciences, public health, ehealth, environmental health, occupational health, big data, smart data, machine learning, deep learning, artificial intelligence, epidemiology, health risk assessment, air pollution, agriculture, gait, modeling, bibliometric analysis
Scopus Publications
Scholar Citations
Scholar h-index
Scholar i10-index
Pascal Petit, Sylvain Chamot, Abdallah Al-Salameh, Christophe Cancé, Rachel Desailloud, and Vincent Bonneterre
Elsevier BV
Felix Muehlensiepen, Pascal Petit, Johannes Knitza, Martin Welcker, and Nicolas Vuillerme
Springer Science and Business Media LLC
AbstractTelemedicine (TM) has augmented healthcare by enabling remote consultations, diagnosis, treatment, and monitoring of patients, thereby improving healthcare access and patient outcomes. However, successful adoption of TM depends on user acceptance, which is influenced by technical, socioeconomic, and health-related factors. Leveraging machine learning (ML) to accurately predict these adoption factors can greatly contribute to the effective utilization of TM in healthcare. The objective of the study was to compare 12 ML algorithms for predicting willingness to use TM (TM try) among patients with rheumatic and musculoskeletal diseases (RMDs) and identify key contributing features. We conducted a secondary analysis of RMD patient data from a German nationwide cross-sectional survey. Twelve ML algorithms, including logistic regression, random forest, extreme gradient boosting (XGBoost), and neural network (deep learning) were tested on a subset of the dataset, with the inclusion of only RMD patients who answered “yes” or “no” to TM try. Nested cross-validation was used for each model. The best-performing model was selected based on area under the receiver operator characteristic (AUROC). For the best-performing model, a multinomial/multiclass ML approach was undertaken with the consideration of the three following classes: “yes”, “no”, “do not know/not answered”. Both one-vs-one and one-vs-rest strategies were considered. The feature importance was investigated using Shapley additive explanation (SHAP). A total of 438 RMD patients were included, with 26.5% of them willing to try TM, 40.6% not willing, and 32.9% undecided (missing answer or “do not know answer”). This dataset was used to train and test ML models. The mean accuracy of the 12 ML models ranged from 0.69 to 0.83, while the mean AUROC ranged from 0.79 to 0.90. The XGBoost model produced better results compared with the other models, with a sensitivity of 70%, specificity of 91% and positive predictive value of 84%. The most important predictors of TM try were the possibility that TM services were offered by a rheumatologist, prior TM knowledge, age, self-reported health status, Internet access at home and type of RMD diseases. For instance, for the yes vs. no classification, not wishing that TM services were offered by a rheumatologist, self-reporting a bad health status and being aged 60–69 years directed the model toward not wanting to try TM. By contrast, having Internet access at home and wishing that TM services were offered by a rheumatologist directed toward TM try. Our findings have significant implications for primary care, in particular for healthcare professionals aiming to implement TM effectively in their clinical routine. By understanding the key factors influencing patients' acceptance of TM, such as their expressed desire for TM services provided by a rheumatologist, self-reported health status, availability of home Internet access, and age, healthcare professionals can tailor their strategies to maximize the adoption and utilization of TM, ultimately improving healthcare outcomes for RMD patients. Our findings are of high interest for both clinical and medical teaching practice to fit changing health needs caused by the growing number of complex and chronically ill patients.
Pascal Petit, Elise Gondard, Gérald Gandon, Olivier Moreaud, Mathilde Sauvée, and Vincent Bonneterre
Springer Science and Business Media LLC
Cécile Manaouil, Sylvain Chamot, and Pascal Petit
Elsevier BV
Sylvain Chamot, Abdallah Al-Salameh, Pascal Petit, Vincent Bonneterre, Christophe Cancé, Guillaume Decocq, Agnès Boullier, Karine Braun, and Rachel Desailloud
Elsevier BV
Pascal Petit, Gérald Gandon, Marc Dubuc, Nicolas Vuillerme, and Vincent Bonneterre
Elsevier BV
Felix Muehlensiepen, Pascal Petit, Johannes Knitza, Martin Welcker, and Nicolas Vuillerme
JMIR Publications Inc.
Background Previous studies have demonstrated telemedicine (TM) to be an effective tool to complement rheumatology care and address workforce shortage. With the outbreak of the COVID-19 pandemic, TM experienced a massive upswing. A previous study revealed that physicians’ willingness to use TM and actual use of TM are closely connected to their knowledge of TM. However, it remains unclear which factors are associated with patients’ motivation to use TM. Objective This study aims to identify the factors that determine patients’ willingness to try TM (TM try) and their wish that their rheumatologists offer TM services (TM wish). Methods We conducted a secondary analysis of data from a German nationwide cross-sectional survey among patients with rheumatic and musculoskeletal disease (RMD). Bayesian univariate and multivariate logistic regression analyses were applied to the data to determine which factors were associated with TM try and TM wish. The predictor variables (covariates) studied individually included sociodemographic factors (eg, age and sex) and health characteristics (eg, disease type and health status). All the variables positively or negatively associated with TM try or TM wish in the univariate analyses were then considered for the Bayesian model averaging analysis after a selection based on the variance inflation factor (≤2.5). All the analyses were stratified by sex. Results Of the total 102 variables, 59 (57.8%) and 45 (44.1%) variables were found to be positively or negatively associated (region of practical equivalence ≤5%) with TM try and TM wish, respectively. A total of 16 and 8 determinant factors were identified for TM try and TM wish, respectively. Wishing that TM services were offered by rheumatologists, having internet access at home, residing 5 to 10 km away from the general practitioner’s office, owning an electronic device, and being aged 40 to 60 years were among the factors positively associated with TM try and TM wish. By contrast, not yet being diagnosed with an RMD, having no prior knowledge of TM, having a bad health status, living in a rural area, not documenting one’s health status, not owning an electronic device, and being aged 60 to 80 years were negatively associated with TM try and TM wish. Conclusions Our results suggest that health status, knowledge, age, and access to technical equipment and infrastructure influence the motivation of patients with RMD to use telehealth services. In particular, older patients with RMD living in rural areas, who could likely benefit from using TM, are currently not motivated to use TM and seem to need additional TM support.
Pascal Petit, Gérald Gandon, Stéphan Chabardès, and Vincent Bonneterre
Wiley
The etiology of central nervous system (CNS) tumors is complex and involves many suspected risk factors. Scientific evidence remains insufficient, in particular in the agricultural field. The goal of our study was to investigate associations between agricultural activities and CNS tumors in the entire French farm manager workforce using data from the TRACTOR project. The TRACTOR project hold a large administrative health database covering the entire French agricultural workforce, over the period 2002‐2016, on the whole French metropolitan territory. Associations were estimated for 26 activities and CNS tumors using Cox proportional hazards model, with time to first CNS tumor insurance declaration as the underlying timescale, adjusting for sex, age and geographical area. There were 1017 cases among 1 036 069 farm managers, including 317 meningiomas and 479 gliomas. Associations varied with tumor types, sex and types of crop and animal farming. Analyses showed several increased risks of CNS tumors, in particular for animal farming. The main increases in risk were observed for meningioma in mixed dairy and cow farming (hazard ratio [HR] = 1.75, 95% confidence interval [CI]: 1.09‐2.81) and glioma in pig farming (HR = 2.28, 95% CI: 1.37‐3.80). Our study brings new insights on the association of a wide range of agricultural activities and CNS tumor and subtype‐specific risks in farm managers. Although these findings need to be corroborated in further studies and should be interpreted cautiously, they could have implications for enhancing CNS tumor surveillance in agriculture.
Felix Muehlensiepen, Pascal Petit, Johannes Knitza, Martin Welcker, and Nicolas Vuillerme
JMIR Publications Inc.
Background Previous studies have demonstrated telemedicine (TM) to be an effective tool to complement rheumatology care and address workforce shortage. With the outbreak of the SARS-CoV-2 pandemic, TM experienced a massive upswing. However, in rheumatology care, the use of TM stagnated again shortly thereafter. Consequently, the factors associated with physicians’ willingness to use TM (TM willingness) and actual use of TM (TM use) need to be thoroughly investigated. Objective This study aimed to identify the factors that determine TM use and TM willingness among German general practitioners and rheumatologists. Methods We conducted a secondary analysis of data from a German nationwide cross-sectional survey with general practitioners and rheumatologists. Bayesian univariate and multivariate logistic regression analyses were applied to the data to determine which factors were associated with TM use and TM willingness. The predictor variables (covariates) that were studied individually included sociodemographic factors (eg, age and sex), work characteristics (eg, practice location and medical specialty), and self-assessed knowledge of TM. All the variables positively and negatively associated with TM use and TM willingness in the univariate analysis were then considered for Bayesian model averaging analysis after a selection based on the variance inflation factor (≤2.5). All analyses were stratified by sex. Results Univariate analysis revealed that out of 83 variables, 36 (43%) and 34 (41%) variables were positively or negatively associated (region of practical equivalence≤5%) with TM use and TM willingness, respectively. The Bayesian model averaging analysis allowed us to identify 13 and 17 factors of TM use and TM willingness, respectively. Among these factors, being female, having very poor knowledge of TM, treating <500 patients per quarter, and not being willing to use TM were negatively associated with TM use, whereas having good knowledge of TM and treating >1000 patients per quarter were positively associated with TM use. In addition, being aged 51 to 60 years, thinking that TM is not important for current and future work, and not currently using TM were negatively associated with TM willingness, whereas owning a smart device and working in an urban area were positively associated with TM willingness. Conclusions The results point to the close connection between health care professionals’ knowledge of TM and actual TM use. These results lend support to the integration of digital competencies into medical education as well as hands-on training for health care professionals. Incentive programs for physicians aged >50 years and practicing in rural areas could further encourage TM willingness.
Pascal Petit
Elsevier BV
Marie-Laure Aix, Pascal Petit, and Dominique J. Bicout
Elsevier BV
Mélodie Valière, Pascal Petit, Renaud Persoons, Christine Demeilliers, and Anne Maître
Elsevier BV
Pascal Petit, Delphine Bosson-Rieutort, Charlotte Maugard, Elise Gondard, Damien Ozenfant, Nadia Joubert, Olivier François, and Vincent Bonneterre
Oxford University Press (OUP)
Abstract Objectives A vast data mining project called ‘TRACking and moniToring Occupational Risks in agriculture’ (TRACTOR) was initiated in 2017 to investigate work-related health events among the entire French agricultural workforce. The goal of this work is to present the TRACTOR project, the challenges faced during its implementation, to discuss its strengths and limitations and to address its potential impact for health surveillance. Methods Three routinely collected administrative health databases from the National Health Insurance Fund for Agricultural Workers and Farmers (MSA) were made available for the TRACTOR project. Data management was required to properly clean and prepare the data before linking together all available databases. Results After removing few missing and aberrant data (4.6% values), all available databases were fully linked together. The TRACTOR project is an exhaustive database of agricultural workforce (active and retired) from 2002 to 2016, with around 10.5 million individuals including seasonal workers and farm managers. From 2012 to 2016, a total of 6 906 290 individuals were recorded. Half of these individuals were active and 46% had at least one health event (e.g. declared chronic disease, reimbursed drug prescription) during this 5-year period. Conclusions The assembled MSA databases available in the TRACTOR project are regularly updated and represent a promising and unprecedent dataset for data mining analysis dedicated to the early identification of current and emerging work-related illnesses and hypothesis generation. As a result, this project could help building a prospective integrated health surveillance system for the benefit of agricultural workers.
Pascal Petit and Dominique J. Bicout
Elsevier BV
Pascal Petit, Anne Maître, and Dominique J. Bicout
Elsevier BV
Renaud Persoons, Laure Roseau, Pascal Petit, Claire Hograindleur, Sarah Montlevier, Marie Marques, Gabriel Ottoni, and Anne Maitre
Elsevier BV
Pascal Petit, Anne Maître, Renaud Persoons, and Dominique J. Bicout
Elsevier BV
Anne Maitre, Pascal Petit, Marie Marques, Claire Hervé, Sarah Montlevier, Renaud Persoons, and Dominique J. Bicout
Elsevier BV
Pascal Petit, Dominique J. Bicout, Renaud Persoons, Vincent Bonneterre, Damien Barbeau, and Anne Maître
Oxford University Press (OUP)
Background
Similar exposure groups (SEGs) are needed to reliably assess occupational exposures and health risks. However, the construction of SEGs can turn out to be rather challenging because of the multifactorial variability of exposures.
Objectives
The objective of this study is to put forward a semi-empirical approach developed to construct and implement a SEG database for exposure assessments. An occupational database of airborne levels of polycyclic aromatic hydrocarbons (PAHs) was used as an illustrative and working example.
Methods
The approach that was developed consisted of four steps. The first three steps addressed the construction and implementation of the occupational Exporisq-HAP database (E-HAP). E-HAP was structured into three hierarchical levels of exposure groups, each of which was based on exposure determinants, along 16 dimensions that represented the sampled PAHs. A fourth step was implemented to identify and generate SEGs using the geometric standard deviation (GSD) of PAH concentrations.
Results
E-HAP was restructured into 16 (for 16 sampled PAHs) 3 × 3 matrices: three hierarchical levels of description versus three degrees of dispersion, which included low (the SEG database: GSD ≤ 3), medium (3 < GSD ≤ 6), and high (GSD > 6). Benzo[a]pyrene (BaP) was the least dispersed particulate PAH with 41.5% of groups that could be considered as SEGs, 48.5% of groups of medium dispersion, and only 8% with high dispersion. These results were comparable for BaP, BaP equivalent toxic, or the sum of all carcinogenic PAHs but were different when individual gaseous PAHs or ∑PAHG were chosen.
Conclusion
Within the framework of risk assessment, such an approach, based on groundwork studies, allows for both the construction of an SEG database and the identification of exposure groups that require improvements in either the description level or the homogeneity degree toward SEG.
Pascal Petit, Anne Maître, Renaud Persoons, and Dominique J. Bicout
Elsevier BV